Voice data entry

When your hands are full, when you cannot get the phone close enough to a target, or when you simply want to collect values in rapid succession, speaking is often faster than aiming the camera. QCR Scanner treats the microphone as a data source alongside the camera: you say a value, the app turns it into text, runs it through your validation rules, and adds it to the active list. Speech recognition runs on your device’s own speech engine, and every value it hears passes through the same rules and the same smart-extraction logic as the values you scan with the camera.

This page explains how to choose between the three voice modes — push-to-talk, continuous listening, and voice verification — when each one is useful, and how voice ties into validation, smart extraction, and list logic.

The three voice modes

QCR Scanner has three distinct voice modes. They all share the same speech-recognition engine; what differs is how they process what they hear.

Mode	What it does	Best for
Push-to-talk (voice)	Listens for a single utterance, transcribes it, validates it, and adds it	One-at-a-time, controlled entry
Continuous listening (continuous)	Listens without stopping and auto-splits speech by digit count	Reading many numbers/serials in a row
Voice verification (verify)	Asks you to read out the value OCR found and confirms the match	Hands-free confirmation of a scan

Push-to-talk (voice)

Push-to-talk is the simplest mode. You tap the microphone button, the app listens for a single utterance, transcribes it, and runs the result through your validation rules. If the value passes all active rules, it is added to the active list; if not, it is skipped. Because you tap the button for each value, you keep full control over when and what you say — stray speech in between never leaks into the list.

This mode is ideal when you want to enter values one at a time in a controlled way, or whenever you need to think, look, or swap the object between readings.

Continuous listening (continuous)

Continuous listening keeps the microphone open and automatically splits the incoming text as you speak. The split is driven by digit count: each run of digits that reaches the expected length is treated as one completed value, validated, and added to the list, after which the app keeps listening for the next value. This lets you read out many serial numbers or figures back to back without tapping the button again and again.

Tip: Continuous mode is most efficient for long lists of same-length numbers — a whole box of serial numbers, for example. Say the numbers at a steady pace with clear gaps between them, and the app will separate each run on its own.

Note: Because continuous listening keeps the microphone open, incidental speech can be evaluated too. Validation rules act as a safety net here: only runs that match your rules enter the list, and everything else is ignored.

Voice verification (verify)

Voice verification uses the camera and your voice together. You first scan a value with the viewfinder; when OCR finds a result, the app waits for you to read that value aloud. It compares what you say against the recognized value:

• If they match, you get green feedback and a confirming vibration, and the value is added with confidence.
• If they do not match, you get red feedback and a warning vibration, so you can look again and correct it.

This mode is useful when you want to confirm that OCR read a value correctly without fixing your eyes on the screen: you read the value out loud, and the phone tells you with a vibration whether it is green or red. On dense or small text — serial numbers, strings of digits — it adds a second layer of checking. For the highest scanning accuracy, you can pair verification mode with the Quality capture mode described on the Scanning page.

How to choose a mode

The voice mode is selected from the microphone control on the scan screen. Your choice determines how each tap of the microphone behaves for that session. You can switch between modes at any time — for example, read a long list in continuous mode, then confirm a suspicious value with verification mode.

Microphone handling runs through a native bridge so that the app’s own speech prompts do not clash with the device’s system sounds: while the app is listening it temporarily mutes the system’s recognition beeps and unmutes them again when listening stops. This happens automatically and needs no configuration.

Microphone permission

Because voice entry accesses the device microphone, it requires Android’s RECORD_AUDIO permission. The first time you start listening in a voice mode, the Android system permission dialog appears; recognition will not run until you grant it. Once granted, it is remembered on your device, like the camera and other permissions.

Note: The microphone is only used while a voice mode is actively listening. For the full list of permissions the app uses, see Privacy.

How it interacts with validation and smart extract

Values that arrive by voice go through exactly the same pipeline as values from the camera. That means all the logic you have set up applies to voice input automatically:

• Validation rules — Every recognized utterance is evaluated against your active regex rules. Only values that pass all rules are added, which keeps unwanted speech out — especially in continuous and verification modes. See Validation rules.
• Smart extract — If a single phrase you speak contains several valid parts, smart extract can break the text into tokens and pull out the ones that pass validation. See Smart extract.
• List triggers — If a value added by voice matches a list-trigger pattern, the active list can switch automatically, just as it does for values scanned with the camera. See Lists.

Voice input is not a separate record path; it is simply another way to add values to the same list. The source of an added value is marked as voice in the record, so you can later tell — in CSV or in sync — whether a value came from the camera or from speech.

Languages

Because speech recognition uses your device’s speech engine, the languages it can hear depend on what is installed on your device. The app listens according to the language active on the system, so you can usually speak in whatever language your phone is set to. The interface itself is offered in 22 languages and is selected automatically from the device language — see Languages for details.

Tip: When reading numbers and serial numbers, you get the best results by speaking the digits separately and clearly. In noisy environments, validation rules will discard misheard values; even so, bring the phone a little closer to your mouth while speaking if you can.

QCR Scanner is made by ReviseTouch.

Next steps

• Validation rules — define which rules voice values must pass
• Smart extract — pull several valid values out of a single utterance
• Scanning — combine verification mode with the Snap and Quality capture modes
• Lists — route voice-added values between lists automatically